Skip to content

[mlir][tensor] Fix computation of collapse in dynamic cases #127560

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

[mlir][tensor] Fix computation of collapse in dynamic cases #127560

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
27 changes: 13 additions & 14 deletions mlir/lib/Dialect/Utils/ReshapeOpsUtils.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,27 +49,26 @@ mlir::getReassociationIndicesForCollapse(ArrayRef<int64_t> sourceShape,
int64_t currTargetShape = targetShape[targetDim];
while (sourceDim < (sourceShape.size() - 1) &&
sourceShape[sourceDim] != ShapedType::kDynamic &&
prodOfCollapsedDims * sourceShape[sourceDim] < currTargetShape) {
(currTargetShape == ShapedType::kDynamic ||
prodOfCollapsedDims * sourceShape[sourceDim] < currTargetShape)) {
prodOfCollapsedDims *= sourceShape[sourceDim];
currIndices.push_back(sourceDim++);
}

if (sourceDim >= sourceShape.size())
return std::nullopt;

// If the current expanded dimension is dynamic, then the collapsed
// dimensions should also be dynamic and product of all previous unprocessed
// dimensions of the expanded shape should be 1.
if (sourceShape[sourceDim] == ShapedType::kDynamic &&
(currTargetShape != ShapedType::kDynamic || prodOfCollapsedDims != 1))
return std::nullopt;

// If the collapsed dim is dynamic, the current expanded dim should also
// be dynamic.
if (currTargetShape == ShapedType::kDynamic &&
sourceShape[sourceDim] != ShapedType::kDynamic)
currTargetShape != ShapedType::kDynamic)
return std::nullopt;

// For static shapes, if the product of dimensions of the expanded shape
// should match the collapsed dimension shape.
if (prodOfCollapsedDims * sourceShape[sourceDim] != currTargetShape)
if (sourceShape[sourceDim] != ShapedType::kDynamic &&
prodOfCollapsedDims * sourceShape[sourceDim] != currTargetShape)
return std::nullopt;

currIndices.push_back(sourceDim++);
Expand Down Expand Up @@ -315,11 +314,11 @@ SmallVector<Range> SliceFromCollapseHelper::getExtractSliceParams(
// have proven that these are not sliced. In this case we just take
// the full extent of each dimension in the reassociation list.
if (linearizedDimensions[it.index()]) {
llvm::append_range(
offsetsSizesAndStrides,
llvm::map_range(it.value(), [&](int64_t idx) -> Range {
return {zeroAttr, collapseShapeInputShape[idx], oneAttr};
}));
llvm::append_range(offsetsSizesAndStrides,
llvm::map_range(it.value(), [&](int64_t idx) -> Range {
return {zeroAttr, collapseShapeInputShape[idx],
oneAttr};
}));
continue;
}

Expand Down
14 changes: 14 additions & 0 deletions mlir/test/Dialect/Tensor/canonicalize.mlir
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1191,6 +1191,20 @@ func.func @compose_expand_of_collapse_dynamic(%arg0 : tensor<4x?x10x64x2xf16>, %

// -----

func.func @compose_expand_of_collapse_dynamic_collapse(%arg0 : tensor<4x13x10x64x?xf16>, %arg1 : index) -> tensor<4x13x10x?xf16> {
%collapsed = tensor.collapse_shape %arg0 [[0, 1], [2], [3, 4]] : tensor<4x13x10x64x?xf16> into tensor<52x10x?xf16>
%expanded = tensor.expand_shape %collapsed [[0, 1], [2], [3]] output_shape [4, 13, 10, %arg1] : tensor<52x10x?xf16> into tensor<4x13x10x?xf16>
return %expanded : tensor<4x13x10x?xf16>
}

// CHECK-LABEL: func @compose_expand_of_collapse_dynamic_collapse
// CHECK-SAME: %[[ARG0:.+]]: tensor<4x13x10x64x?xf16>
// CHECK: %[[RESULT:.+]] = tensor.collapse_shape %[[ARG0]]
// CHECK-SAME: [0], [1], [2], [3, 4]
// CHECK: return %[[RESULT]]

// -----

// CHECK-LABEL: func @zero_rank_reshape_multi
func.func @zero_rank_reshape_multi(%arg0: tensor<f32>) -> tensor<f32> {
// CHECK: return %arg0
Expand Down